Estimates of Prevalence for Autoimmune Disease

First Paper – 1997:  9 million people in the US

The first attempt to find the number of people affected by autoimmune disease was made in 1997 when Denise Jacobson and others published “Epidemiology and Estimated Population Burden of Selected Autoimmune Diseases in the United States." This paper studied 24 diseases, and the estimated number of persons affected by those autoimmune diseases in the US was about 9 million.

Update in 2012: 14.7 million people in the US

In 2012, Drs. Scott Hayter and Mathew Cook published a study, “Updated assessment of the prevalence, spectrum and case definition of autoimmune disease." This study covered 81 diseases, and estimated prevalence in the US at about 14.7 million (see chart below). 

The other numbers you’ve heard:

There are 2 other numbers for autoimmune disease prevalence circulating on the internet:

23.5 million

This number appears in the Report on Autoimmune Diseases prepared by the National Institutes of Health in 2005. The report states that US prevalence for all autoimmune diseases is between 14.7 million and 23.5 million. The lower number is close to the estimate by Hayter & Cook. The high end of the range, 23.5, was an estimate by Dr. Noel Rose, who wrote the book on autoimmune disease (literally – it’s called “The Autoimmune Diseases” and it’s in its 5th printing).  Dr. Rose based this number on the work by Jacobson and took into account potential misdiagnosis, under-diagnosis and the potential that more autoimmune diseases might be discovered. 

The NIH paper does not say which diseases are included in that estimate, but it makes a reference to “80 illnesses”. This also matches closely with the Hayter paper (which was not published until 2012). 

50 million

This number appears on the American Autoimmune Related Diseases Association (AARDA) website.  Although AARDA references the NIH paper, it states that number was based on 24 diseases (that’s the Jacobson list of diseases, not the 80 that the NIH used) and they determined that there are a lot more diseases and extended the statistics to the remaining 120 or so diseases that could be autoimmune.  While we agree with AARDA’s list of diseases, we believe that the diseases missing from the NIH list are the rarer diseases, and we believe the AARDA number overstates the number of people with these diseases.

Over 50 million?

And estimates of “over 50 million” that can be found elsewhere? Well, a lot of people think bigger is better and are trying to draw attention to the problem. We agree there’s a problem, but exaggerating a problem does not solve it. 

Does it matter what the number is?

Some people think that making the number bigger is going to get more attention for this under-recognized class of diseases. But that attention wanes. After a while, as the number does not change, people will get the feeling that it does not matter what they do.

ARI wants a number that can be reasonably measured, repeatedly, over time. IF we can do that, here are a few of the things we can learn:

  1. Is the number of people with autoimmune disease actually increasing or decreasing?

  2. Which diseases are creating the greatest burden on patient health?

  3. If a patient has one autoimmune disease, what other diseases are they at risk for?

ARI believes that good statistics are essential for measuring progress and one of our goals is to find a scientific answer to the question, "How many people in the US have at least one autoimmune disease."

What’s ARI’s approach?

ARI decided that the existing studies have one obvious problem: Not one of them actually tried to count how many patients had autoimmune disease.  Their work is good, but it was limited to essentially studying other studies.

ARI has an idea: how about counting people with disease? We are working with 4 academic medical centers to scour their electronic medical records and see how many people suffer from each of the diseases. This work is underway and it will have problems. We will not be able to account for misdiagnosis and under-diagnosis. Not all diseases have medical codes, and medical codes are often wrong. But it’s a start. And by doing the same study across 4 hospital systems, we can develop statistics that are reasonably consistent for the US population.

The info below shows the data from the study by Hayter & Cook. Note that epidemiologists do not use “Percent of population” in their work. Prevalence statistics are stated as “Cases per 100,000 persons”.  US Prevalence is based on a US population of 320 million.