Everybody Hurts: Old Chronic Pain Data Don’t Improve With Age
A seemingly comprehensive article used as the basis for big claims about the number of people in chronic pain is rooted in data more than a decade old.
It was written in 2003 by Christa Harstall and Maria Ospina in the newsletter Clinical Pain Updates. The data in the article range from 1991 to 2001. Most of the 13 studies predate 2000.
Why does this matter?
For one, serious research into the epidemiology of pain is a relatively young field where more data are constantly being added to the evidence base. This article is part of that attempt to add to the evidence base and, as it describes, it found only 13 studies since 1990 that attempt to measure pain prevalence.
The findings from these 13 studies are all over the place. One of the biggest accomplishments of Harstall’s piece is that it provides a nice table of every study reviewed and shows the prevalence estimate, the number of people who were found to be in pain and the total number of people surveyed. You can see the full table here, but I’m also going to quickly run through the numbers by year and country.
1991, United Kingdom, 11.5%, 1,037
1993: Sweden, 55.2%, 1,609
1993: United Kingdom, 35%, 1,340
1997: Australia, 50.2%, 990
1997: United Kingdom, 13%, 1,953
1998: Canada, 44.4%, 410
1998: France, 32.9%, 741
1998: Global, 21.5%, 5,438
1999: Scotland, 50.4%, 3,605
2000: Netherlands, 25%, 5,423
2000: Israel, 10.1%, 2,210
2001: Australia, 18.5%, 17,496
2002: Spain, 23.4%, 5,000
The differences in the findings stem from different demographic factors, different ways of surveying people, different definitions of pain, and different analytical methodologies. Some use a postal questionnaire. Some use telephone surveys. Some use face-to-face interviews.
In some studies, the definition of pain is feeling pain most days for more than three months. In others the definition is “current and persistent pain most of the time for six months or more during the prior year.”
Even with all of these differences in study design, there are two undeniable trends:
1. All the biggest studies show lower prevalence estimates. Every study that surveyed more than 5,000 people arrived at a prevalence estimate of 25% or less. The largest study, which captured 17,496 people, estimated prevalence of 18.5%.
2. All the most recent studies show lower prevalence estimates. There were a total of 47,252 people surveyed across these 13 studies. The bulk of them – 64% – were surveyed for studies published since 2000. If you look just at those studies, you will see an average prevalence estimate of 19%.
So when the journal Pain Medicine published an article in 2011 claiming that 75 million to 150 million Americans were in chronic pain, the basis for that upper estimate was quite old and, based on the newer data, quite weak.
As the authors of the Scotland study that is the foundation for that 150 million estimate told me, making the statistical leap from 2,000 people in Scotland to 150 million in the U.S. is hazardous. And authors using the Scotland study at any point in the last 10 years had at least four more recent studies to choose from, including three studies that surveyed many more people.
There have been other studies since then, but this is the paper that was cited for the estimate that 75 million to 150 million people are in chronic pain, even though better evidence appears to have been available.
If you just take the estimate from the largest study in that list – 18.5% – and you choose to make the same statistical leap, in this case from Australia to the United States, you could more authoritatively say something like this: “According to prevalence estimates from the largest study in the world – conducted in Australia – an estimated 44 million adult Americans may be in chronic pain.”
That still sounds like a serious problem worth research and policy attention. But it loses the alarming ring of 1 out of every 2 people being in pain.
Photo credit: Nennie T via Flickr